Reading is one of the most compelling ways to experience a story. But the actual act of reading can tell a tale all its own — a neurological story about the reader’s brain.

One doctor has turned a simple reading test into a powerful brain-mapping tool to detect and recommend treatment for dyslexia. Revealing the neurological data underlying reading disabilities, Dr. Ioannis M. Aslanides is changing the story of dyslexia, proving it’s a very misunderstood condition. 

“Reading is a way to map the brain,” said Aslanides, founder of RADAR, a company dedicated to better understanding dyslexia through data.

As an opthalmologist, Aslanides has studied the range of eye conditions. But when his 13-year-old son was diagnosed with marginal dyslexia, he wondered how a man of his profession missed all the signs. With such a late-stage diagnosis, his son would face even more obstacles overcoming the effects of dyslexia.

Aslanides’ son isn’t alone when it comes to the late diagnosis of dyslexia. According to a recent survey published by Made by Dyslexia, a non-profit spearheaded by Sir Richard Branson, almost 28% of survey respondents stated that it took schools more than three years to identify if a student is dyslexic.

Meanwhile, about 23% of respondents stated that in some cases the condition is never identified. Furthermore, if a student does get diagnosed with dyslexia, almost 12% responded that the school did not offer any kind of support to help the student.

If left unchecked, kids with dyslexia often suffer from low self-esteem, behavioral problems, aggression, anxiety and feelings of isolation in terms of socials skills. They may also find it hard to advance their careers when they become adults.

Aslanides set out to find a smarter way of helping educators and institutions diagnose and remedy dyslexia. Enlisting the help of nine other doctors and professionals, the group established a data-driven deep dive into the world of dyslexia, seeking ways to arm educators with this data to customize treatment plans. He hopes to have every student, starting in his home country of Greece, tested for dyslexia multiple times throughout their childhood, beginning as early as possible. 

Intelligent machines help detect dyslexia

RADAR was built by a team of ophthalmologists, speech pathologists, and engineers, backed by decades of research on dyslexia and other learning disabilities. The system uses a computer to test readers with its proprietary software, utilizing two sets of cameras with monochrome CMOS image sensors to track the movements of the eyes. Because eye movements are being tracked, the person being tested doesn’t have to read aloud. RADAR also assesses how long the eyes fixate and how often they revisit letters, words or phrases.

The cameras are placed between the screen and the person, with a down view of the participant’s face. Such a position reduces reflections or shadows for better accuracy. The cameras are focused on one eye each and captures pupil and corneal or CR coordinates, while Eigenfaces software is used to compute the pupil/CR positions.

To begin, the subject is placed in front of a computer screen. There is a chin rest, similar to those used by ophthalmologists, to limit head movements or tremors that can affect the result of the test. When the person is settled in the right position, calibration begins.

There are 25 spots shown consecutively, strategically placed to accurately calibrate the eyes’ full range of movements. The participant is asked to focus on displayed spots until enough fixations are obtained in that region.

After the calibration, the pupil and CR coordinates on the camera images are mapped to the coordinates of the spots where the subject was looking at. Once mapping is done, testing can now proceed.

“If you have a negative RADAR [score], you have a high negative predictive value. That tells you a student has no problem with reading. If a test gives a positive radar, they have problems with reading and likely have a neurological problem. This incorporates more than dyslexia,” Aslanides said.

RADAR uses two types of parameters to determine whether a person has dyslexia: non-word-specific and word specific. Non-word-specific focuses on fixation duration, saccade lengths, short refixations and the total number of fixations during the evaluation. As for word-specific, RADAR assesses gaze duration for each word at the first encounter and the number of revisits. The calculated score is able to quickly determine if the individual has a high probability of being dyslexic.

RADAR boasts of 95% accuracy, boosted by the objective assessment from its mathematical computations. And thanks to modern cloud computing, assessments can be run quickly and cheaply on a global scale.

Distributing such a test via cloud addresses one of the major challenges for dyslexia testing — the cost. Per patient, a dyslexia test could cost upwards of £1,000 ($1,200) in the United Kingdom’s private sector, according to Aslanides. In the United States, the costs more than double, running up to $5,000 per patient. Then there are the committee meetings to review and finalize the assessment, which personally cost Aslanides more than $2,000 for his son’s official diagnosis. 

Dyslexia doesn’t affect just young children, Aslanides cautioned. “There’s a large number of kids entering freshman year at university every year in the U.K., and 40% are marginally dyslexic. It costs the university, because they’re required to provide infrastructure for therapy, assessment, extra exam time,” Aslanides stated. 

Part of the problem, in Greece and beyond, is the lack of official stats and assessment standards when it comes to dyslexia. With some 50 million students in America’s public education systems alone, and an average cost for each at $13,000 per year, allocating funds for the estimated 12% of dyslexic students can run a high tally.

With better data, RADAR hopes to show the many neurological facets of dyslexia, busting the longstanding myths holding entire societies back. While working as a professor in China, Aslanides had the opportunity to dispel the notion that the icon-based, rather than letter-based, Chinese alphabet prevents any chance of dyslexia.

“I found 11% of students were dyslexic, which indicates the problem is neurological,” Aslanides said. “It has nothing to do with the text or the language. It’s ubiquitous.” 

The challenge: standards of education

The biggest challenge for RADAR is delivering the data goods in order to change the status quo within the scientific community, as well as education systems and even governments. With few standards regarding the definition and diagnosis of dyslexia, it’s been something more like guesswork to determine who is dyslexic and how to best treat them. 

In traditional medicine, simply diagnosing someone isn’t sufficient — there must be test data to justify the diagnosis and guide the assessment.  “I could do a blood test and say, ‘You have anemia,’ but that’s not enough,” Aslanides said. “I need to provide quantifiable numbers to contextualize your disorder. [RADAR] is an MRI of reading. We can accurately pinpoint where the problem lies with very high accuracy at the level of one syllable.”

Supporting RADAR’s efforts to acquire useful data is their application of proprietary statistical learning algorithms to quantify the reading level of each subject. 

“When I learned of the existence of the RADAR development in my hometown of Heraklion, I decided to pay the team a visit,” said Joseph Sifakis, the 2007 winner of the esteemed Turing Award and the founder of Verimag, a leading research center specializing in embedded systems. “I’ll admit, I was pleasantly surprised to find the level of science and serious mathematics the team is capable of.” 

After further visits with RADAR administrators, Sifakis believes RADAR’s use of strict mathematical models and artificial intelligence will be life-changing, calling it a “potent tool” for helping students impeded by dyslexia and other reading abnormalities.

The benefits of artificial intelligence are leading several education initiatives around the world, hoping to accelerate students that may otherwise be left behind. A 2019 report by HolonIQ sees the education sector as one in dire need of digital transformation, giving it the opportunity to “leapfrog the typical linear technology path most other industries have followed,” the report reads. 

“This opens the door for education technology, such as AI, voice interfaces, machine learning and others, to address the student cost, access and return on investment challenges and institutional efficiency, effectiveness and growth objectives,” the report adds.

Curating the countermeasures to dyslexia

The data provided to teachers and professional is meant to aid them, not replace them. What’s done with that data could make all the difference in a global society where image-centric social media outlets and voice-activated digital assistants seemingly make it more difficult to recognize the neurological issues underlying dyslexia. 

RADAR is looking to strike a balance between man and machine, uncovering efficiencies through the coupling of distributed cloud applications and powerful algorithms that learn alongside each other. It’s now up to RADAR to establish a global standard to support its goal of streamlining dyslexia detection and curating countermeasures.

Photo: John Jennings/Unsplash